Last March's report is different this March
An auditor asks the team to reproduce a revenue-by-customer-tier figure from a report filed a year ago. The pipeline is deterministic, the raw data is retained, the code is at the same version — and the recomputed figure does not match the filed one.
What you would do first
Answer before revealing anything. The value of the exercise is entirely in committing to a diagnosis you can be wrong about.
- 1Compare the totals first. A matching total with different breakdowns points immediately at an attribute that has moved rather than at rows that have.
- 2Identify which dimension attributes the model groups by, and check whether each is Type 1 or Type 2.
- 3Count how many entities changed the attribute between the report date and now — that count should approximately explain the movement.
- 4Establish whether the change history exists anywhere: retained change records from capture, a snapshot table, or nothing at all.
What is actually going on
The trap
The fix that looks right. Read it even if you got the answer — especially then.
Start snapshotting the dimension nightly from today and call the problem solved. Future reports become reproducible, the filed figure still cannot be reproduced, and the snapshot silently misses every attribute change that happens and reverses between two snapshots.